Accelerating Safe Reinforcement Learning with Constraint-mismatched Policies.
Tsung-Yen YangJustinian RoscaKarthik NarasimhanPeter J. RamadgePublished in: CoRR (2020)
Keyphrases
- reinforcement learning
- optimal policy
- markov decision process
- policy search
- control policies
- markov decision processes
- fitted q iteration
- reinforcement learning agents
- function approximation
- cooperative multi agent systems
- hierarchical reinforcement learning
- partially observable markov decision processes
- reward function
- machine learning
- markov decision problems
- reinforcement learning algorithms
- learning algorithm
- multi agent
- state space
- long run
- data sets
- total reward
- decision problems
- macro actions
- learning agents
- temporal difference learning
- action selection
- state abstraction
- supervised learning
- approximate policy iteration
- learning process
- mobile robot
- continuous state
- control policy
- function approximators
- finite state
- average cost
- linear constraints
- infinite horizon